INDEX

Explanations

phrases and terms related to correctness and verification in coding or computational contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

test consistency and correctness

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

principalColumn

-0.51

 invokingState

-0.48

GOTREF

-0.47

 ProtoMessage

-0.45

DockStyle

-0.44

\{\\

-0.42

uxxxx

-0.42

 namanya

-0.42

MessageWindow

-0.41

featureID

-0.41

POSITIVE LOGITS

 correct

0.53

 ביצ

0.50

 proper

0.48

 consistency

0.45

correct

0.45

Correct

0.44

 accurate

0.41

正确的

0.41

Спољашње

0.40

 Correct

0.40

Activations Density 0.018%

phrases and terms related to correctness and verification in coding or computational contexts

test consistency and correctness

No Comments

No Known Activations